Grasping complexity: analysing risk for sustainable development

(1)

LUND UNIVERSITY PO Box 117 221 00 Lund +46 46-222 00 00

Grasping complexity: analysing risk for sustainable development

Becker, Per

2010

Link to publication

Citation for published version (APA):

Becker, P. (2010). Grasping complexity: analysing risk for sustainable development. Lund University.

Total number of authors: 1

General rights

Unless other specific re-use rights are stated the following general rights apply:

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research.

• You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal

Read more about Creative commons licenses: https://creativecommons.org/licenses/ Take down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

(2)

Grasping complexity: analysing risk for

sustainable development

Per Becker

Department of Fire Safety Engineering and Systems Safety

Lund University

Doctoral thesis

(3)

Grasping complexity: analysing risk for sustainable development Per Becker Report 1047 ISSN: 1402-3504 ISRN: LUTVDG/TVBB—1047—SE ISBN: 978-91-7473-048-7 Number of pages:XX

Illustrations: Per Becker

Keywords: Development, sustainable development, risk, risk analysis, sustainability science, disaster, value, hazard, vulnerability, capability, system, systems approach, human-environment systems, design, design science.

Abstract: Sustainable development relies on our ability to make decisions today that will determine our tomorrow. Given that uncertainty is explicitly allowed to influence our view of what the future holds for us, most ex ante analyses of challenges for sustainable development can be viewed as analysing risk. Many frameworks for analysing risk exist today, but analysing risk for sustainable development entails different requirements. By applying a combination of traditional science and design science, this thesis presents six such requirements, informed by available theory and new empirical studies. The thesis also presents six criteria for scientifically developing frameworks for analysing risk for sustainable development.

LUCRAM (Lund University Centre for Risk Assessment and Management) © Copyright: Per Becker and Department of Fire Safety Engineering and Systems Safety, Lund University, Lund 2010.

Avdelningen för brandteknik och riskhantering

Lunds tekniska högskola Lunds universitet Box 118 211 00 Lund brand@brand.lth.se www.brand.lth.se Telefon: 046-222 73 60 Telefax: 046-222 4612

Department of Fire Safety Engineering and Systems Safety

Lund University P.O. Box 118 SE-211 00 Lund Sweden brand@brand.lth.se www.brand.lth.se Telephone: +46 46-222 73 60 Telefax: +46 46-222 4612 224

(4)

Summary

Sustainable development relies on our ability to make decisions today that will determine our tomorrow. Given that uncertainty is explicitly allowed to influence our view of what the future holds for us, most ex ante analyses of challenges for sustainable development can be viewed as analysing risk. Many frameworks for analysing risk exist today, but analysing risk for sustainable development entails different requirements. By applying a combination of traditional science and design science, this thesis presents justifications for six key requirements, informed by available theory and new empirical studies. Although the empirical base for this thesis is facilitating capacity development for managing risk, it may hold broader implications for analysing risk for sustainable development in general. The identified key requirements comprise the ability to: (1) integrate phenomena on various spatial and temporal scales, as well as structural and functional complexity (systemic); (2) accommodate different stakeholder values (multi-value); (3) incorporate a wide range of initiating events that may impact on what stakeholders value (multi-hazard); (4) integrate a multitude of factors and processes that contribute to the susceptibility of what stakeholders’ value to the impact of the events (multi-susceptive); (5) involve various stakeholders across functional, administrative and geographical borders (multi-stakeholder); and (6) integrate several risk analyses performed by different groups of stakeholders (multi-analysis).

The thesis also presents six criteria for scientifically developing frameworks for analysing risk for sustainable development, namely a systematic and transparent design process in which: (a) the empirical and normative statements behind the framework’s purpose and required functions are explicitly justified and stated; (b) the actual form of the developed framework makes it possible to utilise in practice; (c) the connections between purpose, functions and form of the framework are clear; (d) the framework is utilised in its intended contexts; (e) the utility of the framework is measured in how well its form fulfils the required

(5)

Sammanfattning

Hållbar utveckling bygger på vår förmåga att fatta beslut i dag som avgör vår morgondag. Givet att osäkerhet tillåts påverka vår syn på vad framtiden har i sitt sköte, kan de flesta framåtblickande analyser av utmaningar för hållbar utveckling anses vara riskanalyser. Det existerar många ramverk för riskanalys idag, men att analysera risk för hållbar utveckling innebär särskilda krav.

Genom att tillämpa traditionell vetenskap och designvetenskap, presenteras i denna avhandling motiveringar för sex krav på riskanalyser för hållbar utveckling, baserade på tillgänglig teori och nya empiriska studier. Även om den empiriska grunden för denna avhandling är att underlätta utveckling av kapacitet för att hantera risk, kan dess resultat ha betydelse för att analysera risk för hållbar utveckling i allmänhet. Kraven är förmåga att: (1) integrera fenomen på olika skalor i rum och tid, samt strukturell och funktionell komplexitet (systemisk), (2) tillgodose olika värden (multi-värde), (3) innehålla många inledande händelser som kan påverka dessa värden (multi-hot), (4) integrera en mängd faktorer och processer som bidrar till sårbarheten hos dessa värden (multi-sårbarhet), (5) involvera olika intressenter över funktionella, administrativa och geografiska gränser (multi-intressent) och (6) integrera flera riskanalyser som utförs av olika grupper av intressenter (multi-analys).

Avhandlingen presenterar också sex kriterier för att vetenskapligt utveckla ramverk för att analysera risk för hållbar utveckling, nämligen en systematisk och öppen designprocess där: (a) de empiriska och normativa antaganden bakom syfte och funktion uttryckligen motiveras, (b) den konkreta utformningen av det utvecklade ramverket gör det möjligt att använda det i praktiken, (c) sambanden mellan syfte, funktion och form är tydligt, (d) ramverket används i dess avsedda kontext, (e) användbarheten av ramverket mäts i hur väl dess form uppfyller de funktioner som krävs för att uppfylla sitt syfte, och (f) resultatet av utvärdering efter användning vägleder ytterligare utveckling.

(6)

Acknowledgements

Although I have spent my entire adult life dealing with risk or disaster-related issues in relation to sustainable development and pondering how to grasp their complexity, it was not until I started working at Lund University Centre for Risk Assessment and Management (LUCRAM) that I felt that I was getting somewhere. I am thus grateful to all my colleagues who have assisted me in this immense learning experience, both within Lund University and elsewhere. Especial thanks are due to my Head of Department, Robert Jönsson, for making it administratively possible to finalise a PhD while working as a Visiting Professor, and to my supervisor Professor Kurt Petersen for guiding me along the way. However, the biggest debt I owe is to Henrik Tehler, my co-supervisor and colleague, who through his great intellect and patience has supported the development of my understanding of risk, as well as my own intellectual development in general. Finally, I extend all my love to my wife, Anna, for always supporting me, and to my son, Noa, for inspiring me to continue to be as curious as he is.

Lund, October 18, 2010 Per Becker

(7)

(8)

Table of contents

1. Introduction

“Unless we change direction, we are likely to end up where we are going” - Chinese proverb

This thesis is an attempt to facilitate sustainable development by articulating an argument for the need to integrate the notion and management of risk in development policy and practice. And more specifically by outlining initial ideas for a framework for analysing risk in this context, including potential negative impacts of climate change. 1.1. Background

The world economy has been estimated to have increased around 50 times and the population almost six times from the industrial revolution to the end of our last century (Maddison 2001:28). This development continues to place increasing strains on the world’s natural resources and environment (Kalas 2000; Grimble et al. 2002; Komatsuzaki & Ohta 2007; Syvitski 2008; Gadda & Gasparatos 2009; Fan & Qi 2010), while vast inequalities persist and even deepen both between and within states (Rist 2006:18; Bywaters 2009; Gorringe et al. 2009; O'Brien et al. 2009a). Although the last century saw a global increase in life expectancy (Riley 2001) and a decrease in child mortality (Ahmad et al. 2000:1175) and adult illiteracy (Parris & Kates 2003:8070-8071), the economic development was highly unequal rendering the same wealth in the final decade of the century to the richest one percent in the world as to the poorest 57 percent (Milanovic 2002:50). In order to reduce poverty while striving towards a more viable use of natural resources, it is vital to make future development more sustainable.

Regardless of whether one focuses on economic growth or on more human-centred parameters, such as increased literacy or reduced child mortality, most uses of the concept of development have one thing in common. This is the fact that they project some sort of scenario into the future, in which the variables of interest develop over time along a preferred expected course. This scenario is, in modern society, not believed to be predestined or predetermined in any way, but is dependent on a wide range of human activity, environmental processes,

(11)

2

continuously creating a multitude of possible futures (Japp & Kusche 2008:80), causing uncertainty as to what real development will materialise.

Being unable to see into the future, as well as being largely incapable of predicting it (Simon 1990:7-8; Taleb 2007/2008), modern individuals, organisations and societies resort to the notion of risk in order to make sense of their uncertain world (Zinn 2008:3-10). Risk is a contested concept, but to be able to talk about risk at all entails some kind of idea of uncertain futures as well as of their potential impacts on what human beings value (Renn 1998a:51). This use of risk also entails that risk must be defined in relation to some preferred expected outcome (Kaplan & Garrick 1981; Luhmann 1995:307-310; Kaplan 1997; Kaplan et al. 2001; Johansson & Jönsson 2007:12-14; Zinn 2008:4). If risk is related to potential deviations from a preferred expected future, stakeholders in development must endeavour to reduce such risk to safeguard their development objectives.

There are many courses of events and their underlying processes that may negatively impact development, in either the short or the long term. Abrupt changes in political leadership, global financial crises, algal bloom, epidemic outbreak, droughts, cyclones and outbreaks of communal violence are just a few examples of initiating events that may set off destructive courses of events. Behind these often dramatic courses of events lay processes of change which are less sensational, but may have far-reaching indirect impacts, such as globalisation (Beck 1999; Yusuf 2003; Murad & Mazumder 2009), demographic and socio-economic processes (Wisner et al. 2004:62-74; Satterthwaite et al. 2009:11-19), modernisation (Beck 1992), environmental degradation (Geist & Lambin 2004; Pimentel 2006; Lewis 2006), the increasing complexity of modern society (Perrow 1999b; Perrow 2008), the development of protracted low intensity armed conflicts (Kaldor 1999), and increased asymmetrical threats (Kegley 2003). In addition we have the mounting threats of climate change, not only potentially increasing the frequency and intensity of destructive extreme weather events (Webster et al. 2005; Nordhaus 2006; Syvitski 2008; von Storch & Woth 2008; Elsner et al.

(12)

1 Introduction

2008; Gravelle & Mimura 2008; Kasei et al. 2010), but also changing everyday life for vast numbers of people.

These courses of events and their underlying processes rarely exist in isolation, neither from each other nor from the development activities and processes that they impact. It is thus not only vital to ensure that development gains are durable in the face of destructive courses of events and their underlying processes, but also that the means to reach the development gains do not augment, or create new, risks that hinder development for future generations (WCED 1987). Analysing risk is thus a requisite for sustainable development (Haimes 2004:101-106). There are many frameworks for analysing risk that have been developed over the last four decades or so (e.g. Haimes 1998; Aven 2003). However, analysing risk in a sustainability science context entails additional, and sometimes different, requirements.

The world is increasingly complex (OECD 2003:33-50; Calvano & John 2004:25-26; Renn 2008:5). Facilitating sustainable development requires the ability to integrate phenomena on a wide range of spatial and temporal scales, from local to global and from delayed to immediate (Kates et al. 2001:641). It also demands the ability to grasp structural and functional complexity (ibid.), which means not focusing on individual elements of the world in isolation but on how they are connected, interact with and depend on each other (Haimes 1998:104; Turner et al. 2003a:8077). This causes the consequences of an initiating event to propagate through the system (Rinaldi et al. 2001; Hollenstein et al. 2002:56-61; OECD 2003:44-45; Jiang & Haimes 2004:1215-1229; Dobson et al. 2007). To facilitate sustainable development, societies must have the capacity to manage a wide range of risks (Haimes 1992:415; Haimes 2004:101-106) to a complex set of elements that human beings value. It is vital to include a multitude of initiating events in the analysis and an even larger set of interdependent factors and processes, both social and biophysical (Kasperson & Kasperson 1996:96; Turner et al. 2003a), contributing to the susceptibility of these elements to the direct or indirect impact of the events. It is also vital to include a wide range of stakeholders (Renn et al. 1997:218-219; Haimes

(13)

4

social, political and economical contexts (Renn 2008:8-9), as well as experts, policymakers and the public at large (Renn 2001).

(14)

2 Research demarcation

2. Research demarcation

“The most beautiful part of every picture is the frame” - Gilbert Keith Chesterton (1909/2008:105)

2.1. Presenting the context of the research

Sustainable development is, as will be presented later in this thesis, both conceptually and practically a broad and multifaceted issue (WCED 1987; Kates et al. 2001). It is an issue of paramount importance for the continued existence of the world as we know it. At its core lies the idea that in planning for the future, we must think about what to do and not to do today, in order to bring about that future (Simon 1990:11). An important part of sustainable development is, in other words, forward-looking. However, there may be many ways to envisage the future. A major distinction among these approaches is the extent to which uncertainty is explicitly allowed to influence the resulting view of what the future holds for us. This thesis is limited to concerning itself with frameworks for ex ante analyses of challenges for sustainable development that explicitly include uncertainty (Figure 1), i.e. frameworks for analysing risk for sustainable development.

Figure 1. Demarcation of research context.

Analysing risk for sustainable development entails a broad societal focus, which limits this thesis to excluding analysing risk in more monomorphous contexts, e.g. restricted to one stakeholder (a company), one particular site (a chemical plant), one functional sector (power distribution), etc. The frameworks for analysing risk for sustainable

(15)

6

stakeholders with interest of society as a whole, e.g. national, provincial or local governing bodies, public-private partnerships for regional development, civil-society organisations with broad social responsibility, etc.

2.2. Purpose, research question and process

The purpose of this thesis is to outline a framework for analysing risk for sustainable development, including negative impacts of climate change. To meet this purpose, the thesis intends to answer the following overall research question:

What criteria should guide the design of a framework for analysing risk for sustainable development, and how should such a framework be developed scientifically?

The scientific design of a framework necessitates a somewhat different approach than traditional science (March & Smith 1995; Abrahamsson 2009:20). Instead of being mainly concerned with the pursuit of knowledge (Weber 1949; Ravetz 1996; Checkland 1999:50), the focus must be placed on designing a framework that must meet some predefined purpose (Simon 1996:4-5, 114; Poser 1998:85-87; Cook & Ferris 2007:173; Abrahamsson 2009:20; Hassel 2010:14-15).

The normative focus of this endeavour poses a different challenge than for traditional descriptive research, as normative statements are inferred from value preference and not from empirical observation, a philosophical assumption presented in detail in Section 3.1.3 Axiological assumptions. This challenge opens up the way for an infinite number of possible frameworks that could be considered to meet the stated purpose (Figure 2) (Simon 1996:119-120; Poser 1998:86). Just as it is unfeasible to identify all possible frameworks, it is also unfeasible to design the optimal framework (Simon 1996:119-120; Poser 1998:86; Hevner et al. 2004:88-89). The aim must instead be to design a framework that satisfies some predetermined design criteria (Simon 1996:119-121; Abrahamsson 2009:23; Hassel 2010:40).

To scientifically develop the framework, we must ensure transparency of both what underlies decisions about design criteria and of the design

(16)

process itself, so that they are open to scientific scrutiny (Abrahamsson 2009:22-24; Hassel 2010:42-47). Each decision about a specific design criterion may or may not have implications for other criteria, but the process to establish them can be seen as additive in the sense that each decision determines the path to take through this part of the design process (Figure 2). The set of design criteria is then what the framework is evaluated against.

The research process of this thesis is to establish design criteria, to develop the framework, apply it in context and evaluate it against the established design criteria. This may seem like a rather linear process, but it is inherently iterative (Hevner et al. 2004:88). The design process is presented in full in Chapter 4.2 The design process.

The framework for analysing risk for sustainable development that is outlined in this thesis is only applied and evaluated in one context. This may seem to limit the effectiveness and usability of the framework itself, as more applications most certainly are necessary to guide further development. However, the overall research question of this thesis is not focused directly on the outcome of the design process, which is the first embryo of a framework. Rather, it focuses instead on making a transparent argument for design criteria that should guide the design of such a framework, as well for how such framework can be designed scientifically. Such a focus is less limited by the few applications of the framework itself.

Figure 2. The additive process of

(17)

8 2.3. Appended papers

This thesis is based on a synthesis of six peer-reviewed journal articles. The first five (I-V) are descriptive in nature and are intended to empirically inform the normative argumentation behind the establishment of key design criteria. This connection between the descriptive and the normative is elaborated on in Section 3.1.3 Axiological assumptions. The argumentation behind such criteria is also informed by available theory. The last paper (VI) is used to present initial ideas for a framework for analysing risk for sustainable development and to discern if the used design process can be scientifically rigorous. The setting for the studies described in the appended papers is international and their focus is on facilitating capacity development for managing risk or actual destructive courses of events. This represents one of many possible settings within the context of this research, but still, it forms the empirical basis for indicating potential generalisations in the final chapter of this thesis. The research contribution and implications of the papers are summarised in the table below (Table 1).

2.4. Outline of the thesis

This thesis consists of six chapters, a bibliography and six appended papers. Here follows a brief synopsis of the four remaining chapters: Chapter 3. Philosophical and theoretical framework: This chapter starts by presenting vital assumptions about ontology, epistemology, axiology and complexity. The chapter continues by presenting a theoretical framework for informing the argumentation for key design criteria. Chapter 4. Methodological issues: This chapter presents traditional science and design science as complementary, and equally essential, elements for the purpose of this thesis. It also presents the design process and similarities and differences between research methodologies and methods used during the thesis research.

Chapter 5. Research contribution: The first part of this chapter presents empirical studies which together with established theory are used to inform the argumentation for key design criteria that should guide the design of the target framework. The second part outlines initial ideas for

(18)

a framework for analysing risk for sustainable development, and presents how it is tested and evaluated against the design criteria. The chapter ends by elaborating on, and answering, the overall research question. Chapter 6. Final remarks: The last chapter of the thesis includes a final discussion of implications for analysing risk for sustainable development in general, and presents ideas for future research.

Paper I

Research question

What general results may come from focusing international development cooperation on specific factors influencing disaster risk without acknowledging interdependencies with other factors? Research methods Observation Interviews Research object 4 projects in Tajikistan and Sri Lanka

Research results

Ignoring interdependencies between factors relevant for meeting project purposes results in sub-optimisation problems and in reduced chances for monitoring and evaluation in all four studied projects.

Implications

As it is likely that these problems are general, it is vital to take into consideration interdependencies between factors when planning and implementing projects.

Paper II

What do stakeholders in disaster risk reduction in Fiji express as valuable and important to protect? Research methods Semi-structured interviews Research object 11 stakeholders selected from all levels of state and traditional leadership, important market sectors, and from civil society in Fiji.

There is great variation in what the 11 included stakeholders express as valuable and important to protect.

Implications

As it is likely that there is variation in what stakeholders express as valuable in general, explicit discussions of what is valuable are vital for risk management initiatives, since the lack of such discussions may result in stakeholders pursuing irreconcilable goals.

Paper III

What do groups of municipal and county council civil servants express as valuable and important to protect and what underlies these expressions?

Research methods

Focus groups

Research object

4 focus groups of 7-21 civil-servants each (3 with municipal and 1 with county-council employees) in Sweden.

There is variation in what the four groups of civil-servants express together as valuable in their context, indicating a range of social, cognitive and contextual factors influencing the result.

Implications

Although it is likely that such variation in what groups express together as valuable is a general factor, the picture supplied is likely to be richer than the sum of individual accounts and could also provide a mutual framework for acting together towards common goals.

(19)

10

Paper IV

How do women and men of from three municipalities in El Salvador rank hazards in their communities?

Research methods Structured interviews Research object 69 randomly selected respondents from 3 randomly selected municipalities in El Salvador. Research results

Although there are no statistically significant differences in the ranking of hazards between women and men in the study, variation in livelihood, level of education, locations of dwellings, age, etc, provide statistically significant differences.

Implications

As it is likely that demographic factors may influence how hazards are prioritised in general, a wide inclusion of people is vital for broad public commitment to specific risk management measures.

Paper V

What are the similarities and differences in expressed flows of information and assistance regarding risk and disasters between different administrative levels involved in managing risks and disaster situations in Fiji? Research methods Semi-structured interviews Research object 5 related stakeholders from all administrative levels potentially involved in risk and disaster-related activities in Fiji.

There are substantial discrepancies between what the 5 respondents on different administrative levels express regarding most aspects of the flow of information and assistance between them.

Implications

As there may be considerable differences between what stakeholders express when describing their system for managing risk and disaster situations, it is vital to integrate information from different administrative levels when constructing one comprehensive view of that system.

Paper VI

Purpose

To present justifications for key requirements for analysing risk for sustainable development, and to outline, test and evaluate initial ideas for a framework for analysing risk that meets these requirements.

Research methods

Focus groups Transect walk

Research object

3 focus groups of 7-10 civil-servants and a full day of transect walk.

The paper presents justifications for six key requirements for analysing risk in the context of developing capacities for managing risk for sustainable development (systemic, multi-value, multi-hazard, multi-susceptive, multi-stakeholder, and multi-analysis). It also presents, tests and evaluates, based on ten questions, an initial framework for building human-environment systems and structuring risk scenarios, in terms of their different strengths and weaknesses.

Implications

Although the initial framework seems to meet the six stated requirements to a certain extent, there are still modifications that have to be made and additional applications are necessary. Representing the world as an explicit human-environment system, while involving a multitude of stakeholders, seems central to analysing risk in the complex context of sustainable development.

(20)

3 Philosophical and theoretical framework

3. Philosophical and theoretical framework

“Thoughts without content are empty, ideas without concepts are blind1_{” - Immanuel Kant}

3.1. Philosophical assumptions

3.1.1. Ontological assumption

The world is dynamic and complex (Dewey 1922). Although parts of the world are determined by processes over which human action has little influence, e.g. tidal cycles or the movement of tectonic plates, human activity has increasingly become the most important determining factor of our future (Simon 1996:2-3). However, regardless of what our world is determined by, this thesis rests on the assumption that the world does exist no matter if I am around to observe it or not (Keat & Urry 1975; Blaikie 1991:121). Such realist ontology entails a distinction between the empirical, the actual, and the real domains, where the first covers our experiences of events through observation, the second covers events whether observed or not, and the third covers the real processes that generate events (ibid.). The world is in other words there, but not directly available to us (Hammersley 1992:69).

3.1.2. Epistemological assumptions

In a world that is not accessible through direct observation, where the empirical rests on our experiences of interaction with it, there can be no objective search for truth (Kuhn 1970). In other words, meaning cannot be discovered, but rather, must be constructed through social processes in which there is a constant struggle over what is considered to be true or false (Winther Jørgensen & Phillips 1999:11-12). The philosophical base for knowledge rests on social practice, and on the practical knowledge2_{of people acting and utilising artefacts in specific social}

contexts (ibid.:11-14). This argument follows John Dewey’s pragmatic philosophy in which thought and action never can be separated (Dewey

1_{Author’s translation from German “Gedanken ohne Inhalt sind leer, Anschauungen ohne} Begriffe sind blind” (Kant 1787/1968:75).

(21)

12

1927/1991), as meaning and knowledge are forged in action (Dewey 1906:306-307). His philosophy of knowledge is also inclusive in the sense that everyone is a capable participant in generating knowledge (Greenwood & Levin 2007:61). But what then is knowledge?

Frank P. Ramsey, another pragmatic philosopher, distinguishes knowledge from belief by stating that belief is knowledge only if it is formed in a reliable process and never leads to mistakes (Ramsey 1931/2001:258). It is thus inadequate to believe something, regardless of the amount of empirical support for it, if that belief leads to errors (Sahlin 1990:4). This is much in line with Popper’s idea of falsification in which a theory ceases to be theory when falsified through empirical observation (Popper 2002b). Falsification is, in this approach, the only genuine way for testing theory (Popper 2002a:48), which means that empirical observation may, at best, be consistent with a theory but can never prove it. Knowledge thus comprises beliefs in whose validity we are reasonably confident (Hammersley 1992:50).

Ramsey looks upon theory as being divided into existing entities (here, α, β, γ), axioms and a dictionary, which can be expressed as “(∈α, β, γ) : dictionary . axioms” (Ramsey 1929/1990:131) and is referred to as the “Ramsey Sentence” (e.g. Mellor 1980; Sahlin 1990). Here, the entities are the building blocks of theory, the axioms are the rules for how the entities function and interact with each other, and the dictionary is our ability to find the entities and axioms in the empirical domain of the world.

The “Ramsey Sentence” not only helps us to understand the relationship between the theoretical and the empirical (Sahlin 1990:140-158), but it also gives us a philosophical framework for managing the complexity in what we perceive when observing the real world. Ramsey (2000) distinguishes between the world, which is home of what we try to explain and understand, and our theoretical construction, which is a tool for making sense of the world. In other words, in order to grasp the complexity of reality, we need to implicitly or explicitly create models of it (Conant & Ashby 1970), or systems, as they commonly are referred to later in this thesis. The vital link between reality and our models is our capability to identify what is relevant for what we attempt to explain,

(22)

understand or solve, as well as our capability to test our hypotheses for that particular explanation, understanding and solution.

3.1.3. Axiological assumptions

Researchers involved in traditional science should primarily be concerned with the pursuit of knowledge (Weber 1949; Ravetz 1996), and be as neutral as they possibly can in relation to values (Hammersley 2000:12). Although it is impossible to be objective and value-neutral, according to the epistemological assumptions above, we are not destined to only produce subjective accounts from which only political processes can distinguish the successes from the failures (Blaikie 2000:56). For researchers to be as value-neutral as they possibly can, it is essential to strive to be reflexive and to identify, and get beyond, prejudices and biases. Value-neutrality is, in other words, an unreachable vision or ideal that researchers involved in the pursuit of knowledge must chase with great strength and stamina to be able to get as close as possible to the realities under study (Hammersley 2000:17-18).

The vision of value-neutrality is, however, only vital in the pursuit of knowledge itself. It is not required to attempt to abandon values when it comes to what is perceived to be significant in selecting the areas of inquiry (Keat 1981:38-58). Nor are normative values prohibited from being involved in the process of utilising results from traditional science in solving real-world problems. This may also be done with scientific rigour and is the focus of the complementary design science (Lee 2007:44; Wieringa 2009:2).

Traditional science is well equipped to deal with how things are in the world (Checkland & Holwell 2007:3-5), but less so in dealing with how things ought to be (Simon 1996:5). This distinction between “is” and “ought to be”, i.e. between the descriptive and the normative, has been problematic for scientists for centuries, as it is easy to stray over from the former to the latter if proper care is not taken. David Hume was the first to point out this problem and some scholars claim that Hume advocates a complete division between “is” and “ought to be”, which is rather theatrically illustrated by the principle’s common epithet “Hume’s

(23)

14

Statements about “how things ought to be” cannot be empirically inferred from statements about “how things are”, as these two are entirely different from each other (Hume 1739/1978:469). For instance, the normative statement “the authorities should lower the speed limit on all 90 km/h roads to 70 km/h” cannot be empirically inferred from the descriptive statement “the number of fatalities per car in accidents is 38% lower on 70 km/h roads compared to 90 km/h roads”. Descriptive statements should, however, be allowed to inform normative statements, but the statement itself will always be essentially inferred from a value preference, i.e. if we value the potentially saved human lives higher than the costs and inconvenience of longer travel times. It is thus vital to be transparent in what values normative statements rest upon (Hammersley 1992:4).

3.1.4. Reductionism, holism and complexity

Contemporary science is concerned with the pursuit of knowledge of a wide range of phenomena, as well as with solving an equally wide range of problems. How these phenomena and problems should be approached has been debated for decades, mainly in the form of more or less fierce advocacy for two seemingly disparate standpoints; reductionism and holism (e.g. Malanson 1999:746-747).

When looking at this conflict, much appears to depend on the definition of reductionism (van den Bergh & Gowdy 2003:76). If reductionism entails the standpoint that a system is nothing but the sum of its parts (e.g. Polkinghorne 2010), it is irreconcilable with holism. However, if reductionism, instead, entails the idea that to address the whole requires a decomposition of it into its parts and examining relations between these parts, they need not to be fundamentally conflicting (van den Bergh & Gowdy 2003:76).

Reductionist approaches attempt to isolate fundamental parts of a phenomenon or problem, and to examine relations between some of these parts, while assuming others to be constant. Holistic approaches, on the other hand, advocate the view that more complex phenomena or problems as wholes cannot be explained, understood or solved by

(24)

studying parts in isolation1_{(Smuts 1926:86-87). There are at least two}

reasons for this. The first may be deducted from the famous statement that the whole is more than the sum of its parts2_{, indicating that there may} be observed regularities in the whole that cannot be obtained by aggregating the regularities of each part (Holland 1998:225). The second is related to the assumption of ceteris paribus, that all things that are not explicitly studied stay the same over time. This is refuted if relations between parts are non-linear and precludes attempts to hold selected parts constant (Ashby 1957:5; Anderson 1999:217). However, the vastness and sheer complexity of our universe make it impossible to research anything without reducing it to some extent (Churchman 1970:B43-44).

According to this view, the issue is not whether reductionism or holism is the best approach, but instead how much complexity is allowed to be involved in approaches to particular phenomena or problems (van den Bergh & Gowdy 2003:76). The guiding principles for finding a satisfactory approach regarding complexity are found in the Law of Requisite Variety (Ashby 1957:202-268) and Ockham’s Razor3

(Checkland 1999:35-36). The former states that a model of reality can only model something in reality if it has sufficient complexity to represent it, while the latter asserts that this complexity should be kept to a minimum and limited to only what is relevant for each particular phenomenon or problem (ibid.). The decisive feature in relation to how this phenomenon or problem may be approached is thus its complexity. Although complexity is a contested concept with a wide variety of definitions and uses (see Backlund 2002), it is, in this context, commonly related to the number of involved parts; the number, significance and/or non-linearity of relations between parts and/or

1_{It is important to note that in order for any parts to be part of a whole they must coincide in} space and time, and there must be some kind of causal dependence between them (Mellor 2006:140).

(25)

16

heterogeneity in space1_{and time}2_{(e.g. Yates 1978:R201). However, if}

any theory or model of reality is a social construction and influenced by both descriptive and normative aspects, then the complexity of any phenomenon or problem must also be related to factors associated with the observer (Ashby 1973:1; Flood 1987:177-178; Wu 1999:3). Ashby’s analogy of the brain demonstrates this with almost amusing clarity and simplicity:

“To the neurophysiologist, the brain, as a feltwork of fibers and a soup of enzymes, is certainly complex [...]. To a butcher the brain is simple, for he has to distinguish it from only about thirty other « meats »” (Ashby 1973:1).

What is included and what is excluded when addressing a phenomenon or problem are vital considerations (Churchman 1970:B43-44; Midgley et al. 1998:467) which are often referred to as boundary judgements to underline their inherently subjective nature (Ulrich 1996:156-158; Ulrich 2002:41).

Boundary judgements, descriptive statements about what we know about the phenomenon or problem, and normative statements about the purpose, objectives, etc of addressing them, are all connected to each other in such a way that changing one automatically induces changes in the others (Ulrich 2000:251-252). It is thus crucial to scrutinise systematically what is included and excluded when addressing any phenomenon or problem, as well as what descriptive and normative statements on which those boundary judgements are based. This is especially so when there are multiple stakeholders in the process, who may agree neither on the boundary judgements nor on the descriptive and normative statements behind them. This systematic scrutiny of boundary judgements is often referred to as boundary critique (Ulrich 1996:171-176; Midgley et al. 1998:467-470; Ulrich 2000:254-266).

1_{What Yates (1978) refers to as broken symmetry.}

(26)

3.2. Theoretical framework

3.2.1. Development, sustainable development and risk

Although the word development has been used for at least 250 years (Harper 2010), it was not until the end of the Second World War that it became an important concept (Thomas 2000b:3). Ideas about development have changed back and forth since then, e.g. the Soviet model of development (Smekal 1991:32-39), modernisation theories (Rostow 1960; Organski 1965), dependency theory (Dos Santos 1970; Frank 1967/2004), World Systems Theory (Wallerstein 1974), Another Development (Hettne 1995:160-206), Human Development (ul Haq 1995). This has spurred numerous and often competing definitions and has made it difficult to communicate about development. It has been suggested that this Babylonian confusion, to a great extent, is the result of the concept being used in three different ways: (1) as a description of a desired future state of society; (2) as a process of change over time; or (3) as deliberate efforts of various stakeholders aimed at improvement (Thomas 2000a:29).

Development may, in other words, refer to a desired state (goal), the process of getting there (change), as well as our efforts to get there (activities). Presenting a desired state of society implies some variable or set of variables (y) that human beings value and aspire to change from its current state. Development is, in other words, inherently normative (Seers 1969/1989). In this context, the process of change is the transformation of the variable or set of variables over time and our efforts refer to purposeful activities we carry out in order to drive or steer this change towards the desired state. The three parts are thus fundamentally related to each other (Thomas 2000a:29), enabling us to look at development as having five components:

1. A variable or set of variables (y) that human beings value and aspire to change.

2. A descriptive statement about the current state of “y”. 3. A normative statement about the desired state of “y”.

4. A normative description of a preferred expected scenario of change in “y” over time.

(27)

18

Most definitions of development only include one or a few of these components explicitly (e.g. Seers 1969/1989:481; Todaro 1989:620; South Commission 1990:10-11; UNDP 1990:10-11; Chambers 1997; Rist 2006:13). However, these components can be seen as incremental in the sense that it is unfeasible to focus a definition of development on one without at least implicit involvement of the others before it. For example, it is impossible to define a desired state of something in the world without first defining that something and determining its current state that requires development, or to define development activities without expressing this desired state and the required change that the activities are designed to bring about (Örtengren 2003:9-15).

Hettne (1995) argues that development is contextual and therefore eludes any fixed and final definition. Because facilitating sustainable development requires emphasising the importance of human activity, the concept of development in this thesis includes all five components. Development is thus viewed as a preferred expected scenario of change in a variable or set of variables (y) over time from a current to a desired state and includes purposeful activities to drive or steer this change (Figure 3).

Sustainable development is commonly defined as “development that meets the needs of the present without compromising the ability of future generations to meet their own needs“ (WCED 1987:43). Looking more closely into the term “sustainable”, one sees that it is defined as something that is “able to be upheld or defended” (The New Oxford American Dictionary 2005). The first part of this definition indicates that sustainable development is development that can be maintained over time, while the second part indicates that sustainable development is development that can be safeguarded from the impact of negative Figure 3. Development as a preferred expected scenario of change in “y” over

time from a current to a desired state, including purposeful activity.

(28)

courses of events and their underlying processes. These two parts are closely related, as it is not only negative courses of events that may impact development, but the means for development may also increase or create new courses of events and underlying processes that in turn make it difficult to maintain development over time. Hence, sustainable development is development that can be maintained over time and be safeguarded from the impact of negative courses of events and their underlying processes.

An important example of the connection between the two meanings of sustainable development is our dependency on burning fossil fuels for energy. This seems to be the main cause behind the climate change that threatens the development of our society. These threats are manifested both through increasing extreme weather events (Webster et al. 2005; Nordhaus 2006; Syvitski 2008; von Storch & Woth 2008; Elsner et al. 2008; Gravelle & Mimura 2008; Kasei et al. 2010) and through the gradual degradation of entire aquatic

and terrestrial ecosystems, on which society depends (Folke & Rockström 2009; Rockström et al. 2009). Regardless of whether they are sudden and dramatic, or gradual and obscure, negative courses of events and their underlying processes may cause deviations from our preferred expected development scenario (Figure 4), limiting the sustainability of our development.

As the future is uncertain (Japp & Kusche 2008:80) and human beings are fundamentally incapable of predicting it (Simon 1990:7-8; Simon 1996:147-148; Taleb 2007/2008), there is not only one but a multitude of possible scenarios that deviate to various degrees from our preferred expected scenario (Figure 5) (Abrahamsson et al. 2010:22-23; Hassel 2010:29). Human beings have the ability to design their future by structuring these uncertain scenarios and use them as mental tools to Figure 4. Potential deviation from the

(29)

20

activities that appear to lead to our desired state or goal (Simon 1990:11; Renn 2008:1). It is in this context that sustainable development can be viewed as requiring the ability to manage risk (ISDR 2004:18-30; UNDP 2004:9-27).

Risk is a contested concept with numerous definitions, creating the potential for miscommunication and misunderstandings (Fischhoff et al. 1984; Rosa 1998; Aven & Renn 2009a). In everyday language, the term “risk” stands for a destructive incident that may or may not occur (Sjöberg & Thedéen 2003:16). Researchers use the term more precisely, but the exact definition of the concept varies (Nilsson et al. 2000:21; Renn 2008:12-45; Aven & Renn 2009a:1-2). Nevertheless, these definitions have three aspects in common. First they all distinguish between reality and possibility, as the concept of risk makes no sense at all if the future is predetermined or independent of present human activity (Renn 1992:56; Renn 2008:1; Zinn 2008:3-4). The future must, in other words, be uncertain (Renn 1998a:51; Renn 1998b:51; Japp & Kusche 2008:80) and any future event must at least be perceived as being amenable to alteration (Zinn 2008:4), i.e. there would be no risk in gambling if the game were 100 percent rigged. Secondly, all definitions of risk explicitly or implicitly entail that these uncertain futures must have the potential to impact1_{what human beings value}

(Renn 1998a:51; Renn 1998b:51; Renn 2008:2), or at least be so perceived (Slovic et al. 1982; Slovic

1987). In other words, there would be no risk in gambling, even if the game were not rigged, if the stake is a grain of sand and it takes place in a desert. Finally, and closely related to the previous aspect, risk must be defined in relation to a preferred expected outcome (Kaplan & Garrick 1981; Luhmann 1995:307-310; Kaplan 1997; Kaplan et al.

1_{Risk is here assumed to relate to negative outcomes (Renn 2008:2), while opportunity denotes} positive outcomes.

Figure 5. Potential deviations as risk scenarios

(30)

2001; Johansson & Jönsson 2007:12-14; Zinn 2008:4). This means that there would be no risk in gambling, even if the game were not rigged and the stakes were high, if the participant has no preference for winning or losing. Taking these three aspects of risk together means that risk is a representation of potential negative deviations in any variable or set of variables representing what human beings value (y) from its preferred expected development over time (Figure 5), making risk analysis into the practise of structuring unwanted scenarios, risk scenarios, and compare them against the preferred expected scenario. The notion of risk depends, in other words, on values and preference, is socially constructed and does not exist ontologically (Slovic 1992:119; Renn 2008:2-3; Aven & Renn 2009a:8-10). However, what does exist ontologically is the complex combination of events and their underlying processes that determine what actually happens (Aven & Renn 2009a:8-10). The actual course of events that produce consequences that human beings experience, interpret and include when making sense of the present as well as structuring scenarios for envisaging the future. These direct or indirect experiences create a link between risk as a social construction and reality (Renn 2008:2), making it vital not to mix ontology and epistemology (Rosa 2010), since these are entirely different philosophical assumptions.

The realist ontology presented in Chapter 3.1 Philosophical assumptions does not require the epistemological realism of viewing risk as real and objective (Slovic 1992:119; Kunreuther & Slovic 1996:119; Renn 2008:2-3; Aven & Renn 2009a). Nor does the social-constructivist epistemology require an ontological constructivism that reduces risk to only subjective issues of power and interest (Renn 2008:3; Aven & Renn 2009a:9). Instead, the ontology and epistemology presented in this thesis form a philosophical foundation for highlighting this link between risk, as socially constructed, and the real world. It is important to note that all human beings take part in experiencing and interpreting the world, making the social construction of risk rooted equally in science and in public values and preferences (MacGregor & Slovic 2000:49; Renn 2008:3-4; Aven & Renn 2009a:8-9).

(31)

22

This conceptual approach to development, sustainable development and risk indicates that all three concepts are essentially connected to each other. Facilitating sustainable development entails purposeful human activity to make sure that any potential deviation from the preferred expected development scenario is avoided or minimised. Anticipating potential deviations, or risk scenarios, is therefore vital for managing risk. Hence, analysing risk is a requisite for sustainable development (Haimes 2004:101-106).

Since analysing risk, according to this approach, is the practise of structuring risk scenarios and comparing them with the preferred expected scenario, a risk analysis is the answer to three questions (Kaplan & Garrick 1981:13): (1) What can happen?; (2) How likely is it to happen?; and (3) If it happens, what are the consequences? Answering these questions, often referred to as “set of triplets” (Kaplan & Garrick 1981; Kaplan 1997; Kaplan et al. 2001) or “risk triplets” (Kaplan 1982; Garrick 2002), entails a systematic analysis of what human beings value, the initiating events that can have a negative impact on those valued elements, and how susceptible they are to the impact the initiating events.

3.2.2. Values and what is expressed as valuable

Values may be seen as “desirable trans-situational goals, varying in importance, that serve as guiding principles in the life of a person or other social entity” (Schwartz 1994:21). In other words, values are what people care about (Keeney 1992:3). To grasp what human beings value in relation to analysing risk, it is important to understand how values come to be ascribed to whatever is declared to have value.

“No man is an island, entire or itself” (Donne 1624:415). This timeworn quote by a 17-century English poet indicates that human beings are social beings, functioning together in society. Giddens takes this idea further by stating that how human beings experience their social context influence how they perceive and understand it, and therefore also how they will act in that social context (Giddens 1984). These actions in turn produce and reproduce social structures, which guide and restrict the actions human beings may take (ibid.:25-26).

(32)

Human actions are thus fundamentally linked to social structures, which are representations of established patterns of behaviour and have the purpose of keeping order while coordinating stable activities (Hardcastle et al. 2005:224). What human beings value is, in other words, socially constructed in context, where prolonged human action creates social structures that direct human beings in what value to ascribe to objects. It is however rare that society is totally homogenous, granting room for individual variation as there may be several social structures competing for dominance. Indicating that the more heterogeneous the society, the more individual variation is possible in what human beings value. Values may thus be seen as acquired “both through socialization to dominant group values and through the unique learning experience of individuals” (Schwartz 1994:21).

Values are notoriously difficult to measure (Slovic 1995:369) and the methods used are predisposed to biases (Payne et al. 1992:121-122; Hassel et al. 2009:36-37) regardless of the assumptions upon which the value elicitation is based (Fischhoff 1991). However, for the purpose of analysing risk it is not the values themselves that we need to elicit, but what human beings express as valuable and as important to protect. In order to understand what people express in particular contexts, it is important to consider that “we can know more than we can tell” (Polanyi 1966/1997:136). What people know can be divided into explicit knowledge and tacit knowledge (Nonaka 1994:16). Explicit knowledge consists of concepts, information and insights that are possible to specify, store and directly transmit to others (Connell et al. 2003:141). Tacit knowledge, on the other hand, is not directly transmittable and consists of knowledge that makes up our mental models for creating meaning to our experiences, as well as our know-how and skills to apply in specific contexts (Polanyi 1967; Nonaka 1994:16; Polanyi 1966/1997:139-140). Explicit and tacit knowledge are, however, closely connected, as “explicit knowledge must rely on being tacitly understood and applied” (Polanyi 1966:7).

(33)

24

awareness which are not identifiable in isolation (ibid.:2-7). An example of this is the psychiatrist showing his students a patient having a seizure. After letting the students discuss if it was an epileptic or a hystero-epileptic seizure, he settles the argument by stating “you have seen a true epileptic seizure. I cannot tell you how to recognize it; you will learn this by more extensive experience” (Polanyi 1961:458). The statement that the seizure was a true epileptic seizure is possible to transmit across to the classroom and is an example of explicit knowledge. The knowledge that the psychiatrist uses for diagnosing the patient is, however, tacit knowledge and is less straightforward to share with the students. It is then only the diagnose itself that is in his focal awareness and accessible to him, as he is only subsidiary aware of each of the many clues and indicators that he more or less subconsciously had observed.

What is in our focal awareness is not only determined by individual characteristics, knowledge, etc, but is constantly changing depending on context. Each situation gives us a sense of what is relevant for what we are doing. Our experience of similar situations, our idea of what the situation calls for or demands, our sense of aim or direction, etc, all combine in supplying us with this “relevance structure” (Marton & Booth 1997:143). What we have talked about recently, what roles the people around us have, what goals we think they have, etc, are thus crucial for how we understand, interpret and remember incoming information. These mental structures or processes are referred to in cognitive science as “schemata” and are constantly amendable (Bartlett 1932/1995:208). The current schemata of an individual guide her interpretation of the incoming situation as well as her expectations of, and attention to, it (Boland et al. 2001:394). Our tacit knowledge comprises, in other words, a part of our schemata (Nonaka 1994:16). Another closely-related cognitive tool that we utilise to get by in our complex world is called “script”. Scripts are cognitive chains of expectations of actions and effects in particular situations (Schank & Abelson 1977; Abelson 1981), which assist individuals in their reaction in those situations without focusing much of their focal awareness on their actions. A main function of schemata and scripts is to facilitate

(34)

coherence in our perception and experience of a situation by filling in gaps in the actual information available.

What stakeholders consider valuable and important to protect is usually not explicitly stated when analysing risk, but instead relies on an implicit assumption that all stakeholders agree on this issue (Nilsson & Becker 2009). The theoretical framework presented makes such assumptions appear dubious at best and flawed at worst. Paper II and III investigate further the validity of such assumptions.

To summarise, what individuals express as valuable in any given situation is socially constructed in context, and is determined by their values and by what they have in their focal awareness at that time. This takes into account the functions of their current schemata and scripts, and it indicates that human beings construct their own mental models of reality through active selection and interpretation of information around them (Vennix 2001:14). It may then be argued that it is of no use to discuss what is valuable and important to protect, as each account is destined to be subjective and fragmented. However, explicit dialogue may facilitate the integration of individual mental models, each of which giving a limited perspective on the world, into one shared model, which is vital for creating a common understanding of the challenge at hand (ibid.). It is probably true that each individual account is unlikely to give a complete picture of what is considered valuable on their own. But it is likely, in a dialogue between several individuals, that what is mentioned triggers additional scripts and amends schemata, thus activating additional knowledge by moving it from their subsidiary awareness to their focal awareness. What the group comes up with is however also highly contextual, but it is still likely to be a richer picture than the sum of each individual account. And more importantly, it is their common picture of what is valuable, making it achievable for the stakeholders involved to formulate and pursue common goals when analysing risk. Without such an explicit common picture, there is a grave danger that the stakeholders might unwittingly impede each others’ work by focusing on protecting different things, e.g. the ministry of agriculture focusing on securing state revenues by promoting the production of cash